2006MNRAS.368..715M

The relationship between galaxies and dark matter (DM) can be characterized by the halo mass of the central galaxy and the fraction of galaxies that are satellites. Here, we present observational constraints from the Sloan Digital Sky Survey on these quantities as a function of r-band luminosity and stellar mass using galaxy-galaxy weak lensing, with a total of 351507 lenses. We use stellar masses derived from spectroscopy and virial halo masses derived from weak gravitational lensing to determine the efficiency with which baryons in the halo of the central galaxy have been converted into stars. We find that an L_*galaxy with a stellar mass of 6x10¹⁰M_☉ is hosted by a halo with mass of 1.4x10¹²h^–1M_☉, independent of morphology, yielding baryon conversion efficiencies of 17⁺¹⁰_–5per cent (early-types) and 16⁺¹⁵_–6per cent (late-types) at the 95 per cent confidence level (statistical, not including systematic uncertainty due to assumption of a universal initial mass function). We find that for a given stellar mass, the halo mass is independent of morphology below M_stellar= 10¹¹M_☉, in contrast to typically a factor of 2 difference in halo mass between ellipticals and spirals at a fixed luminosity. This suggests that stellar mass is a good proxy for halo mass in this range and should be used preferentially whenever a halo mass selected sample is needed. For higher stellar masses, the conversion efficiency is a declining function of stellar mass, and the differences in halo mass between early- and late-types become larger, reflecting the fact that most group and cluster haloes with masses above 10¹³M_☉host ellipticals at the centre, while even the brightest central spirals are hosted by haloes of mass below 10¹³M_☉. We find that the fraction of spirals that are satellites is roughly 10-15 per cent independent of stellar mass or luminosity, while for ellipticals this fraction decreases with stellar mass from 50 per cent at 10¹⁰M_☉ to 10 per cent at 3x10¹¹M_☉or 20 per cent at the maximum luminosity considered. We split the elliptical sample by local density, and find that at a given luminosity there is no difference in the signal on scales below 100h^–1kpc between high- and low-density regions, suggesting that tidal stripping inside large haloes does not remove most of the DM from the early-type satellites. This result is dominated by haloes in the mass range 10¹³-10¹⁴h^–1M_☉, and is an average over all separations from the group or cluster centre.